Australian data center developer Firmus disclosed on Tuesday that it has entered into a multi-year agreement with artificial intelligence powerhouse OpenAI to deliver computing power from two of its Malaysian data centers, positioning the ChatGPT creator as a cornerstone client for the company.
While the financial specifics of the agreement were not disclosed, the strategic importance of data center compute capacity continues to rise amid surging demand for infrastructure capable of running advanced AI models. Following the OpenAI deal, Firmus' total contracted computing capacity across all its clients has climbed to over 900 megawatts (MW).
Traditionally, the AI compute market has been dominated by tech giants wielding substantial capital and scale advantages, enabling them to source cutting-edge GPU resources at lower unit costs. Firmus' business model seeks to challenge this dynamic by deploying infrastructure at scale and leveraging revenue-sharing mechanisms to offer Nvidia's top-tier computing power to smaller AI developers in a more flexible manner.
Firmus' asset portfolio currently includes two operational AI data centers located in Australia and Singapore, with an additional five facilities under development across the Asia-Pacific region. The company plans to deploy Nvidia's next-generation Vera Rubin processors extensively throughout the region, and the expansion into Malaysia further broadens its operational footprint in the area.
Malaysia has emerged as Southeast Asia's fastest-growing data center market, although the rapid pace of facility construction has drawn increased scrutiny over electricity and water resource usage. The agreement arrives amid widespread market speculation that Firmus is preparing for an initial public offering (IPO) this year, which could rank among Australia's largest listings in recent memory.
The AI infrastructure firm was last valued at over $10.5 billion during its most recent funding round, which attracted participation from Nvidia (NVDA.US), Jane Street, funds affiliated with Blackstone (BX.US), and Coatue Management.
In June of this year, Firmus also announced a strategic partnership agreement with Nvidia aimed at delivering cost-effective high-performance computing resources to emerging AI enterprises. Under that arrangement, Firmus will procure Nvidia infrastructure and sell cloud services based on Nvidia technology to clients such as "AI Native." In return, Nvidia receives product revenue along with a share of cloud service income.
According to the terms, Nvidia will deliver 170,000 graphics processing units (GPUs) to a facility on Indonesia's Batam Island between the first quarter of 2027 and early 2028. Firmus projects that the collaboration could generate between $25 billion and $30 billion in revenue over its first six years, underpinned by existing client offtake commitments.
Comments